Paratuberculosis, commonly called Johne's disease, is a chronic infection of the small intestine caused by Mycobacterium avium, ssp. paratuberculosis (“MAP”). Paratuberculosis occurs in a wide variety of animals, but most often in ruminants, especially cattle. The disease presents with symptoms including diarrhea, severe weight loss and decres=ased milk production. Cattle normally become infected with MAP as calves, but because of the slowly progressive nature of the infection, clinical signs of paratuberculosis are usually not seen until animals are adults. There is no cure for the disease and infected animals ultimately become emaciated and must be removed from the herd much sooner than their non-infected counterparts.
Since the signs of paratuberculosis can be confused with the signs of several other diseases, a diagnosis can be confirmed only by use of laboratory tests. The best way to avoid paratuberculosis is to be as certain as possible that animals brought into the herd are not infected with MAP. There are currently three common ways to test animals for paratuberculosis: culture of fecal samples, DNA probe on fecal samples, and blood tests for antibodies to MAP. The fecal culture tests take 8 to 16 weeks because of the extremely slow growth rate of MAP. MAP bacteria can also be detected in fecal samples by use of sophisticated DNA probe tests. DNA probes are much faster than culturing the organism and can be done within three days. Unfortunately, the commercial kit for doing the DNA probe tests are not yet as sensitive as culture and are only able to detect infected animals when their infection has progressed to the stage where large numbers of MAP are being excreted in the feces. Therefore, animals in early stages of the infection are not detected. There are several blood tests for paratuberculosis, but ELISA tests are considered the most accurate and best standardized. Three ELISA-based tests are licensed by the U.S. Department of Agriculture for detection of MAP-infected cattle. The ELISA tests are fast, simple, inexpensive and able to detect animals that are infected before they show signs of paratuberculosis.
However, all of these test results come too late. The animal is already infected. In addition, tests performed on individual animals are not 100% sensitive, meaning they cannot detect 100% of all infected animals. Instead, the tests are performed on a group of animals to extrapolate that if an entire group tests negative, then the probability the group is free of MAP infection is very high.
Methods for paratuberculosis control depend on the type of animal and the patterns of husbandry. In principle, two strategies must be employed at the same time:                1. newborn animals must be protected from infection by being born and raised in a clean environment and fed milk free of MAP; and        2. adult animals carrying the MAP infection must be identified by laboratory tests and removed from the herd, flock or enclosure.        
A national study of US dairies, Dairy NAHMS 96, found that approximately 22% of US dairy farms have at least 10% of the herd infected with paratuberculosis. The study determined that infected herds experience an average loss of $40 per cow in herds with a low paratuberculosis clinical cull rate, while herds with a high paratuberculosis clinical cull rate lost on average of $227 per cow. This loss was due to reduced milk production, early culling, and poor conditioning at culling. The cost of paratuberculosis in beef herds still needs to be determined.
Therefore, there remains a need for methods of predicting animals that have an increased susceptibility of contracting paratuberculosis and selectively breed away from that increased susceptibility. Paratuberculosis is a good candidate for genetic selection because a) an effective vaccine is not available, b) the disease is not curable, c) it causes significant economic losses, and d) it is potentially zoonotic. Selective breeding to reduce disease susceptibility would be a low cost, sustainable practice.
Previous reports of association of DNA markers with paratuberculosis susceptibility have been limited, and frequently focused on candidate genes. The nucleotide-binding oligomerization domain containing 2 gene (NOD2), previously referred to as the caspase recruitment domain 15 protein gene (CARD15), is a well characterized gene that contributes to predisposition to Crohn's disease in humans (see recent reviews by Hugot (2006) and Radford-Smith and Pandeya (2006)) and has been the subject of study in cattle as a candidate gene. Taylor et al. (2006) identified 36 NOD2 polymorphisms in a screening of 42 animals from ten different breeds. Association of these polymorphisms with infection could not be adequately tested owing to a paucity of infected animals (n=11). Subsequently, Pinedo et al. (2009a) tested association of three of the NOD2 polymorphisms identified by Taylor et al. (2006) in a case-control study using cattle of dairy (Holstein, Jersey) and beef (Brahman×Angus) types. An association significant at a nominal P<0.01, after controlling for breed, was found for a non-synonymous SNP in the leucine-rich repeat domain of the gene. Evidence for this association came principally from the Brahman×Angus subset of the data. The same data was subsequently re-analyzed considering effects of predicted SNP haplotypes. A haplotype based on two non-synonymous NOD2 SNPs was found significantly associated with infection status (nominal P<0.0001) in an analysis that did not account for breed. The effect attributable to this risk haplotype was due to greater incidence of infection in animals heterozygous for the haplotype (i.e. overdominance). This is in contrast to the effects associated with NOD2 alleles associated with susceptibility to Crohn's disease in humans where the affects manifest in a partial recessive fashion with genotype relative risk increasing exponentially between risk allele heterozygotes to homozygotes or compound heterozygotes (Economou et al. 2004). Analysis of the NOD2 locus in US Holstein cattle in the author's laboratory (unpublished) revealed additional polymorphisms, but none of nine previously or newly identified SNPs genotyped were significantly associated with infection status in a case-control study using 169 case (positive to either ELISA or fecal culture tests or both) and 188 control cows. In addition, only weak evidence of SNP association with infection status was observed for bovine chromosome 18 (location of NOD2) in whole-genome association analyses reported herein. Pinedo et al. (2009a) point out that the NOD2 allele showing association is more frequent in the Brahman×Angus cattle than in the Holstein cattle they utilized which could account for the lack of association observed in the current work with Holsteins.
Only two whole genome scans for paratuberculosis susceptibility have been previously reported. Our earlier study of three large sire families (264 to 585 daughters per sire) from Population 1 examined 159 informative microsatellite markers across all 29 autosomal chromosomes. One significant (chromosome-wide P-value=0.032) region on chromosome 20 was found, but the wide spacing of the markers made it impossible to more narrowly localize the region (Gonda et al., 2007). Power of this study was lessened by low marker density and the consideration only of linkage effects. The other previously reported whole genome scan utilized the recently available bovine 50 k SNP set to greatly improve marker density. Settles et al. (2009) used 218 Holstein cows in a case-control design to assess marker association with MAP infection under various definitions of infected phenotype. Phenotypes were assigned based on culture of MAP from fecal and tissue samples (ileum, ileo-cecal valve and ileo-cecal lymph nodes). 112 animals were negative to both tests, with the remainder positive to one or both fecal or tissue culture. Composition of case and control groups varied depending on definition of phenotype (fecal-positive vs. fecal-negative, tissue-positive vs. tissue-negative, etc.) leading in some instances to a small number of case samples (range 25-90). Suggestive associations (p<5×10−5) were found under various phenotypic definitions on chromosomes 1, 3, 5, 7, 8, 9, 16, 21 and 23. Correspondence between the results reported here and results reported by Settles et al. (2009) are slight, and none are the specific SNPs that Settles et al. found most significant.
Crohn's disease in humans bears some similarity to Johne's disease in cattle in its manifestation, and as a consequence, genes implicated in the development of Crohn's disease have been considered as candidate genes in the study of Johne's disease. Whole genome association (WGA) studies of Crohn's disease in humans (Barrett et al. 2008; Raelson et al. 2007; Welcome Trust Case Control Consortium 2007; Parkes et al. 2007; Rioux et al. 2007; Libioulle et al. 2007) have been more numerous and of larger scale than the study reported herein. Validated results from human Crohn's disease WGA studies, compilation viewable at www.genome.gov/26525384 (Hindorff et al. 2009), have now implicated more than 30 unique chromosomal regions in humans. The correspondence between results reported here or by Settles et al. (2009) for cattle and the results from humans is limited. Applying an arbitrary and liberal constraint of significant human and bovine markers being within a distance of 4 Mb, only the associations reported by Settles et al. (2009) on proximal BTA9 show correspondence with human WGA results and only associations on BTA7 and 20 reported herein show correspondence. Prostaglandin E receptor 4 (PTGER4) and the immunity-related GTPase family, M gene (IRGM), have been identified as candidate genes for the regions corresponding to BTA7 and 20, respectively in human studies. Regarding PTGER4, Libioulle et al. (2007) identified and validated SNP associations in a 1.25 Mb gene desert on HSA5 adjacent to PTGER4 and found SNP associations with variation in PTGER4 expression. Prior work has found that PTGER4 knock-out mice develop severe colitis upon treatment with dextran sodium sulphate, unlike knock-outs for other prostaglandin receptors (Kabashima et al. 2002) supporting its consideration as a candidate gene. Regarding IRGM, The most significant SNP on BTA7 is located within 2 Mb of the location of IRGM, a candidate gene for Crohn's disease in humans based on results from three whole genome association studies (Barrett et al. 2008, Welcome Trust Case Control Consortium 2007, Parkes et al. 2007) and subsequent studies. The SNPs significantly associated with Crohn's disease in this case flanked the IRGM gene, and subsequent analyses failed to reveal non-synonymous SNPs with the IRGM coding regions leading to speculation that functional polymorphism might alter regulation of IRGM. Subsequent work by McCarroll et al. (2008) identified a 20 kb insertion—deletion polymorphism upstream of IRGM that correlated with differences in IRGM expression, and the authors have speculated that this difference in IRGM expression may related to differences in autophagy.